Anda di halaman 1dari 11

Review Invasive fungal infections

British Society for Medical Mycology proposed

standards of care for patients with invasive
fungal infections
David W Denning, Christopher C Kibbler, and Rosemary A Barnes on behalf of the British Society for Medical

Outcomes for invasive fungal infections have greatly

improved in the past decade, and several new antifungal
drugs have been or will be licensed in the next few years.
Early accurate diagnosis and appropriate treatment have
major impact on survival. In a 1995 survey of laboratory
practice in the UK for mycology, major disparities were seen,
with many laboratories not undertaking even simple
diagnostic procedures. Delays in processing and inadequate
procedures for handling samples, incomplete or delayed
reporting of results, or a combination of these, compromise
the care of patients. In randomised trials of antifungal
chemotherapy, optimum treatments and good alternatives
for others have been defined for some infections. High- Figure 1. Microscopy with Calcafluor white showing C neoformans.

quality care requires a multidisciplinary approach to

diagnosis and management. In this review, we propose diagnosis and treatment of complications of therapy
microbiology, histopathology, radiology, and clinical auditing (including infection). The actual practice of UK
standards, with the evidence base for each reviewed. The microbiology laboratories for mycology is far from ideal.6
standards are absolutes, and, therefore, provide a In view of these issues, the British Society for Medical
straightforward basis for improving services to patients if Mycology set out to define standards of care to lower deaths
they are all implemented. from invasive fungal infections to a minimum. Around
UK£30 million is spent in the hospital sector on antifungal
Lancet Infect Dis 2003; 3: 230–40 treatments,7 not always in an optimum way. Here, we define
some absolute standards of care that can be audited,
The incidence of invasive fungal infections in all developed resulting from extensive consultation. Scales for the quality
nations has increased notably in the past two decades.1–4 All of evidence underpinning these standards are those adopted
over the world the rate of invasive fungal infections by the Infectious Diseases Society of America.8 In all cases a
increased in patients with HIV infection until the advent of category from E to A is used (panel 1), and for therapeutic
combination antiretroviral chemotherapy reversed this interventions a scale of 3–1 (panel 2) has been used.
trend. Diagnosis of some invasive fungal infections is We have set out the standards of care by medical
straightforward, such as most cases of cryptococcal specialty, rather than by disease. The standards are collected
meningitis, but others are problematic, especially invasive together in panels for ease of use in the clinical setting
aspergillosis. Units with active surveillance and diagnostic (panels 3–6).
programmes help to keep mortality to a minimum, whereas
delayed or missed diagnosis inevitably means death. The Microbiology standards of care
differences can be striking, such as in Dijon, France, where All fungi (yeasts and moulds) obtained from sterile sites,
mortality from invasive aspergillosis fell from about 60% to including blood and continuous ambulatory peritoneal
12% over 6 years because of improved management.5 dialysis (CAPD) fluids, and intravenous-line tips should be
The most important improvements in mortality from identified to species level by referral to a specialist laboratory
cancer have come from improved supportive care, including
the management of invasive fungal infection. The UK has a DWD is at the School of Medicine, University of Manchester and
poor outcome from cancer compared with other developed Wythenshawe Hospital, Manchester, UK; CCK is at the Department
countries, and attempts to rectify this situation in the past of Medical Microbiology, Royal Free Hospital, London, UK; and RAB
5 years have been successful only in certain instances. is at the the Department of Microbiology, University Hospital,
Cardiff, UK.
The National Health Service Cancer Plan (http://www.
Correspondence: Dr David W Denning, Education and Research makes no mention of Centre, Wythenshawe Hospital, Manchester M23 9LT, UK.
infection as a major complication in cancer patients, and in Tel +44 (0)161 291 5811; fax +44 (0)161 291 5806;
particular does not address supportive care, including the email

230 THE LANCET Infectious Diseases Vol 3 April 2003

For personal use. Only reproduce with permission from The Lancet Publishing Group.
Invasive fungal infections

Panel 2. Categories indicating the quality of evidence for

Panel 1. Categories indicating the strength of evidence of
recommendations for therapeutic or diagnostic
each recommendation
Category Definition
Grade Definition
A Good evidence to support a recommendation for use
1 Evidence from at least 1 properly randomised controlled trial
B Moderate evidence to support a recommendation for use
2 Evidence from at least 1 well-designed clinical trial without
C Poor evidence to support a recommendation for or against use
randomisation, from cohort or case-controlled analytic studies
D Moderate evidence to support a recommendation against use
(preferably from >1 centre), from multiple time series, or
E Good evidence to support a recommendation against use
dramatic results in uncontrolled experiments
3 Evidence from opinions of respected authorities on the basis of
clinical experience, descriptive studies, or reports of expert
(recommendation A, panel 1), mainly because of organisms’ committees
different susceptibilities to antifungal agents. Bronchoscopy
fluid is deemed sterile in this context for all fungi except
Candida spp. A second reason for species-level identification Fungi cultured from urine of special-care and
is epidemiological, in that identification of particular species transplant patients
can indicate point-source outbreaks. A third reason is that All fungi from urine of patients in intensive care, special-care
the risk of acquisition of an invasive fungal infection or baby units, burn units, and from any transplant patients
interpretation of the result differs depending on the species. should be speciated (A). Candida spp are cultured from
Important examples of fungi that have low susceptibility urine in 1–9% of patients staying in hospital. The
to antifungal agents include: Candida krusei, which is importance of the proportion, from a treatment perspective,
intrinsically resistant to fluconazole9,10 and less susceptible to depends on the clinical context. In infants in special-care
amphotericin B than are other Candida spp;11,12 Aspergillus baby units, this finding frequently represents invasive
terreus13–15 Scedosporium apiospermum (Pseudallescheria infection, and can precede documentation of candidaemia.
boydii), Trichosporon beigelii, and Scopulariopsis spp, which The isolate in the urine is generally identical to that in blood
are resistant to amphotericin B;16 Mucorales spp, which are or at other sites in multiply colonised patients. In other
all intrinsically resistant to all licensed azoles;17 and Candida patients, urine represents an important site of colonisation
glabrata, which is frequently less susceptible to fluconazole that can inform prophylactic and pre-emptive strategies.31
than are other Candida spp.10,16 The most common species encountered are Candida albicans
Several outbreaks of invasive fungal infection have been and C glabrata, but any pathogenic Candida spp and
uncovered because of the unusual species implicated occasionally other fungi, such as Aspergillus spp,
(recommendation B), such as Candida lusitaniae,18 Candida C neoformans, and Coccidioides immitis (as an example of an
tropicalis,19 Candida parapsilosis,20–22 C krusei,23 Candida imported infection), may be involved. The possibility of
lipolytica,24 Pichia anomala,25 among others. Interruption of disseminated disease in patients with such infections is high
transmission in the hospital first requires recognition. and early identification of the most likely infecting species is
Certain species of fungi are more or less likely to cause important.
disease than others (B). Examples include C tropicalis and No evidence shows that higher colony counts in urine
Aspergillus niger. C tropicalis is occasionally found in the are more or less likely to indicate local or systemic disease,
human gut, and is highly likely to become invasive in and isolation from urine derived from an in-dwelling
neutropenic patients and probably in patients in intensive bladder catheter should be viewed in the same light as that
care,26,27 but not in liver-transplant patients.28 Some isolates obtained from spontaneously voided urine. Up to 50% of
are resistant to fluconazole. By contrast, A niger is less likely patients with autopsy-proven invasive candidiasis have
than Aspergillus fumigatus to cause pulmonary disease in negative blood cultures,32 despite multiple blood cultures,
patients with leukaemia; therefore, if it is isolated from but most have positive cultures from other sites, including
bronchoalveolar lavage, it could be colonising but not be the urine. If enough blood is cultured (20 mL for aerobic culture
cause of invasive infection, although this distinction is not is recommended33) new blood-culture systems have a high
absolute.29,30 Some genera of fungi are more difficult to yield, although not 100%.34
identify to species level than others, and some have little
relevance to clinical care. If a reference mycology laboratory Importance of microscopy
cannot identify an isolate to species level, a judgment should Microscopy is an important investigation for several reasons.
be made as to whether the species is required or not. First, the diagnostic yield is more than that for culture alone
Uncertainty surrounds the usefulness of species-level in infections (A). Several studies attest to the substantially
identification for all Candida spp isolates cultured from higher yield with microscopy than that with culture in
respiratory fluids, particularly in patients in whom the bronchoalveolar lavage fluid35–38—in the order of 20% in
isolation is repeated (C). Recognition of Cryptococcus immunocompromised patients. For example, in a study in
neoformans in the context of immunocompromised patients cancer and stem-cell-transplant patients, culture from
is very important, however, and if a laboratory decides not to bronchoalveolar lavage fluid was positive in 40% and
speciate all yeasts isolated from repiratory fluids, a means of cytology in 64%; either or both tests were positive in 67% of
screening for C neoformans needs to be implemented patients.36 In another study, 22 patients with invasive
(figure 1). aspergillosis were diagnosed by bronchoalveolar lavage, and

THE LANCET Infectious Diseases Vol 3 April 2003 231

For personal use. Only reproduce with permission from The Lancet Publishing Group.
Review Invasive fungal infections

damaged by tissue homogenisation and fail to grow. Thus

Panel 3. Microbiology auditing standards in systemic the only means of establishing a causal diagnosis (aside from
fungal disease (strength of recommendation) biopsy or autopsy) is microscopy.
1 All fungi (yeasts and moulds) obtained from sterile sites, including An important differential diagnosis of pneumonia in
blood and CAPD fluids, and intravenous line tips, should be identified immunocompromised patients is Pneumocystis carinii
to species level by referral to a specialised laboratory if necessary (A,
pneumonia. This infection can be diagnosed only by
see panel 1). Bronchoscopy fluid is regarded as sterile in this context
for all fungi except Candida spp microscopy with use of a monoclonal antibody stain, or by
2 All fungi from urine of patients in intensive care, special care baby and
molecular means, since the fungus cannot be cultured on
burn units, and any transplant patients should be speciated (A) standard media. Co-trimoxazole prophylaxis is partly
3 All bronchoscopy fluids from patients suspected of infection should be protective, but resistance is emerging.44
examined microscopically for hyphae and cultured on specialised Specialised fungal media, such as Sabouraud dextrose
media (A) agar provide higher yield in recommendation A
4 All clinical isolates of aspergillus should be identified to species level, infections,45 including a greater frequency of isolation of
by referral to a specialised laboratory if necessary (A) common fungi such as Aspergillus spp, and is the only
5 All cerebrospinal fluid (CSF) specimens from HIV seropositive patients, means (with the possible exception of serology) of making
transplant recipients or patients with sarcoidosis, or CSF specimens
the diagnosis of infection with more fastidious fungi
showing abnormal concentrations of glucose, protein, or leucocytes
without an adequate explanation should be tested for cryptococcal such as Histoplasma capsulatum, C immitis, Penicillium
antigen, if the Gram stain is negative (B) marneffei, and other rare fungi. As international travel
6 All CSF specimens from immunocompromised patients or those with increases, the frequency of such rarities is likely to rise, and
sarcoidosis, or CSF specimens showing concentrations of glucose, it is important to identify which species is implicated to
protein, or leucocytes without an adequate explanation should be guide therapy.
cultured and antigen tested for C neoformans and all bacterial plates
incubated for a minimum of 5 days and fungal plates at 30⬚C for up to
21 days (B) Isolates of Aspergillus spp
All clinical isolates of Aspergillus spp should be identified
to species level, by referral to a specialist laboratory if
in five (23%) microscopy alone was positive.39 Detailed necessary. Aspergillus terreus is resistant to amphotericin
studies in which methods are compared, such as different B.13,14,46 This pathogen used to represent perhaps only 3%
approaches to sample concentration and the importance of of all Aspergillus spp that caused invasive disease, but
different stains, have not been done. We believe that the in a randomised study of voriconazole compared with
sample should be concentrated by centrifugation (B), and amphotericin B for invasive aspergillosis, A terreus caused
the pellet cultured for fungi, mycobacteria, and bacteria and 6% of infections.47 In patients with chronic granulomatous
stained with Gram stain or fluorescent (for chitin) stain for disease, Aspergillus nidulans may be implicated and
recommendation A. The advantages of the fluorescent does not respond as well to amphotericin B as does
stain—eg, those using Blankophor—are the rapidity with A fumigatus.48 Resistance to triazoles (other than
which the stain can be processed and read (<10 s), and fluconazole) in Aspergillus spp has been described only in
improved sensitivity (B; figure 1).40–42 clinical isolates of A fumigatus and not other species.49
The second reason microscopy is important is that it is A niger is proportionately more likely to represent
rapid.38 Results should be available within 2–4 h of the colonisation than infection.30
sample being received by the laboratory. Since delayed
diagnosis of an invasive mould infection of the lung may be Investigation of cerebrospinal fluid
fatal for the patient, rapid processing is important. All cerebrospinal fluid (CSF) samples from HIV-seropositive
Microscopy may help to discern whether an infection is patients, transplant recipients, or patients with sarcoidosis, or
caused by a septate or non-septate mould (figure 2). The
former includes Aspergillus spp, and the latter mucorales
such as Rhizomucor spp, Rhizopus spp, Mucor spp, Absidia
corymbifera, Cunninghamella bertholetiae, and members of
the families Saksenaeaceae, Mortierellaceae, and
Syncephalastraceae. Invasive aspergillosis can be treated with
the azoles itraconazole, voriconazole, or posaconazole
(when licensed), the mucorales can only thus far be treated
with amphotericin B. Pulmonary mucormycosis responds
well to surgical resection (11% mortality vs 68% with
pharmacotherapy in recommendation B), if clinically
possible.43 In addition to an early indication of the best
treatment, pending a positive culture, most cases of
pulmonary infection caused by mucorales do not yield
positive culture. Mucorales are particularly susceptible to
chilling in the refrigerator, and the potential yield may fall Figure 2. Microscopy of bronchoscopy fluid showing non-septate hyphae
with temporary storage of the sample. They can also be indicative of a zygomycete.

232 THE LANCET Infectious Diseases Vol 3 April 2003

For personal use. Only reproduce with permission from The Lancet Publishing Group.
Invasive fungal infections

CSF samples with abnormal concentrations of glucose, media is recommended when the diagnosis is considered. The
protein, or leucocytes without adequate explanation should be diagnosis is not, however, always considered, and
tested for cryptococcal antigen, if the Gram stain is negative C neoformans is sometimes cultured on bacterial media.
(B). Cryptococcal meningitis may be diagnosed reliably by Incubation for at least 5 days for all media plated with CSF
antigen testing (A).50 Most cases of cryptococcal meningitis from patients with abnormal CSF should, therefore, be
occur in HIV-antibody-positive patients (whether previously routine. Fungal media should be incubated at 30⬚C for up to
documented or not), transplant recipients, occasionally in 21 days, the time limitation being drying of plates, which can
other corticosteroid-treated patients, and in patients who have be kept to a minimum with incubation in sloped or water-
sarcoidosis. Antigen testing is appropriate in some patients containing sealed containers (A).
with other immunocompromised states, such as those with
hairy-cell or chronic lymphocytic leukaemia, but these groups Histopathology standards of care
of patients are difficult to define absolutely and, therefore, are Use of fungal stains
omitted from the standard. Rarely, in the UK, cryptococcal All tissues from immunocompromised (including
meningitis occurs in non-immunocompromised patients, in corticosteroid-treated) patients with suspected infection
whom the onset is gradual and commonly manifests as should be stained with fungal stains such as periodic acid-
idiopathic hydrocephalus.51 In immunocompromised patients, Schiff, silver, or fluorescent stains, in parallel with regular
the CSF antigen is positive in more than 95% of cases. Culture stains. Biopsy and surgical resection of abnormal tissue are
is occasionally negative because the sample volume was small frequently definitive features in the diagnosis of invasive
(eg, <5 mL), was not cultured for fungus (although fungal infections. Speed is critically important in achieving an
C neoformans will normally grow on conventional media), or early diagnosis, and the practice of assessing haematoxylin and
the culture was discarded too soon. eosin stains of tissues before deciding whether to use
A diagnostic dilemma arises if a CSF sample is abnormal specialised stains for fungi frequently introduces fatal delays
but no diagnosis is established. Cryptococcal meningitis for patients. Clear indication on all pathology request cards
should be included in the differential diagnosis. In non- that the patient is immunocompromised is critically
immunocompromised patients, culture of large volumes of important (see clinical standard 1, panel 6). Hyphae and yeasts
CSF are sometimes necessary to establish the diagnosis, are commonly invisible on standard sections stained with
whereas the cryptococcal antigen is generally detectable, at low haematoxylin and eosin or Gram stain alone (B).57 Often, only
titres. The ELISA testing format that has been introduced may fragments of hyphae are present, especially on small samples
be the most sensitive method.52 Rarely, false-positive results such as transbronchial biopsy samples. Hyphae are best
occur, almost all in undiluted samples or visualised by specialised stains for fungus (A), as
1-in-2 dilutions. Other fungal causes of meningitis include recommended in anastomotic biopsy samples in lung-
Candida spp, C immitis,53 Aspergillus spp,54 H capsulatum,55 and transplant recipients.58 Some histologists prefer periodic acid-
Sporothrix schenckii,56 all of which are best diagnosed with Schiff staining because the morphology of the tissue adjacent
specialised serology, except Candida spp meningitis, which is to the fungi can be better visualised. Others prefer silver stains
best diagnosed by culture. such as Gomori methenamine silver staining, since hyphae are
CSF samples from immunocompromised patients or immediately visible, but interpretation can be difficult because
those who have sarcoidosis, or abnormal samples showing much besides fungal-cell walls are stained (figure 3).
inexplicably abnormal concentrations of glucose, protein, or Mucorales may require longer staining times, and other
leucocytes should be cultured and antigen tested for fungi can easily be left for too long. Therefore, inclusion of
C neoformans, and all bacterial plates incubated for a good control sections is mandatory. In one study, workers
minimum of 5 days and fungal plates at 30⬚C for 21 days (B). assessed different stains in experimental aspergillus
The diagnosis of cryptococcal meningitis can be elusive; if not keratitis and found that the Gomori methenamine silver
initially suspected on clinical grounds, the diagnosis can be
established only by culture or antigen detection. As a fail-safe,
and because some antigen tests are falsely negative (3–5%),
culture for C neoformans (and other rarer fungi) on fungal

Panel 4. Histopathology auditing standards in systemic

fungal disease (strength of recommendation)
1 All tissues from immunocompromised (including corticosteroid-
treated patients) with suspected infection should be stained with
fungal stains such as Periodic acid-Schiff, silver, or fluorescent (for
chitin) stains, in parallel with regular stains (B). Positive results should
be telephoned through to clinicians immediately (A)
2 Reporting of specimens containing any fungal elements should
always include the presence and absence of yeast forms, hyphae, Figure 3. Histological section of stomach stained with the Grocott
and whether hyphae are or are not septate, if it is possible to tell, and silver stain showing multiple hyphae, which branch irregularly and
whether there is any melanin present (A). The relative size of any fungi demonstrate folding, and appear as septa. The haematoxylin and eosin
should be described, their cellular location, and any specialised stain was negative for fungal elements. Cultures grew the mucorales
structures or forms (A) Absidia corymbifera.

THE LANCET Infectious Diseases Vol 3 April 2003 233

For personal use. Only reproduce with permission from The Lancet Publishing Group.
Review Invasive fungal infections

stain with a haematoxylin and eosin counterstain gave the

best results.59 The key recommendation is, therefore, that
at least one specialised stain for fungi should be used for
tissues from immunocompromised patients concurrently
with haematoxylin and eosin and other relevant stains.
Specialised stains for cryptococcus such as the
mucicarmine stain or the Masson-Fontana stain for
dematiaceous fungi can be done after fungi are seen in
sections. The visualisation of any fungi in sections is
critically important, and the treating clinician must be
informed promptly (A).

Reporting of samples containing any fungal

The reporting of samples containing any fungal elements Figure 4. A large cavitating mass seen on a CT scan of the thorax in a
should always include the presence and absence of yeast man receiving corticosteroids for an autoimmune disorder. Blood culture
and skin biopsy were positive for C neoformans.
forms, hyphae, whether hyphae are septate or aseptate,
and whether any melanin is present. The size of any
fungi, their cellular location and any specialised structures Geotrichum candidum, Acremonium spp, Scopulariopsis
or forms should be described (A). spp, Penicillium spp, Schizophyllum commune, and others.
The appearance of fungi histologically may be sufficient Bulbous ends are typical of Scedosporium apiospermum
to partly establish the cause before culture results (A). The (which is resistant to amphotericin B) and should be
clinical presentation combined with the histological described if seen, although treatment with echinocandins
appearances is frequently sufficient to guide treatment may yield such structures in Aspergillus spp.
appropriately until confirmatory evidence is available. For The finding of small intracellular yeasts is virtually
example Candida spp generally manifest yeasts and hyphae diagnostic of either histoplasmosis or P marneffei
in deep tissue, with the exception of C glabrata, for which infection; the latter can be distinguished from the former
only medium-sized yeasts are visible without hyphae. The by the appearance of one septum separating two halves
hyphae of the mucorales are broad, hyaline, and non- of the yeast, since P marneffei is a fission yeast and
septate, and are commonly seen in necrotic tissue, whereas H capsulatum a budding yeast, and P marneffei stains
those of Aspergillus spp and other moulds are narrower and darker. C neoformans and Blastomyces dermatitidis appear
septate. Folding of hyphae can cause the false appearance of as large yeasts, the former with narrow junctions between
septa and rarely a septum can be seen in mucorales hyphae. cells, the latter with broad-based buds.
Branching patterns differ as well, in that the mucorales C neoformans also has a capsule in most instances,
tend to branch at right angles, at irregular intervals, best seen with the mucicarmine stain. Other rarer fungi
whereas 45⬚ branching is typical of Aspergillus spp. have characteristic histological appearances including
Appearances may differ slightly after antifungal C immitis (spherules) and S schenkii (cigar-shaped
treatment. Several moulds have hyphal forms in tissue bodies). In other diseases, such as mycetoma, grains
indistinguishable from those of Aspergillus spp, including may be seen.
Scedosporium spp, Fusarium spp, Paecilomyces spp,
Radiology standards of care
Panel 5. Radiology auditing standards in systemic fungal Leukaemia, haemopoietic stem-cell transplant, and
disease (strength of recommendation) neutropenic patients
All patients who have leukaemia have undergone
1 All leukaemic and haematopoietic stem-cell-transplant patients who
are or have been profoundly neutropenic (<500 neutrophils/mL) with:
haemopoietic stem-cell transplant, or have recently been
l a new cough, chest pain or haemoptysis severely neutropenic (<500 neutrophils/␮L), who have also
l an abnormal chest radiograph a new cough, chest pain or haemoptysis, an abnormal chest
l a new positive culture of an Aspergillus spp or other mould from radiograph, a new positive culture of an Aspergillus spp or
any site
other mould from any site, microscopic evidence of hyphae
l microscopic evidence of hyphae in any invasive sample
l or unresolved temperature after 7 days of antibiotics and/or
in any invasive sample, or unresolved temperature after
antifungals should have a high resolution (or spiral) computed 7 days of antibiotics, antifungals, or both, should have a
tomography (CT) scan of the chest within 48 h, with immediate high-resolution or spiral computed tomography (CT) scan
consultant review (A) of the chest, ideally within 48 h of changes, and immediate
2 All transplant recipients with a new positive culture of aspergillus or consultant review (A).
other mould should have a CT scan of the chest within 48 h (B)
Imaging plays a crucial part in the diagnosis and
3 All immunocompromised patients with new neurological features (eg, management of pulmonary mould infections, as well as in
change in mental status, seizure, stroke, persistent headache) or
possible or proven meningitis should have a CT or magnetic-
other infections such as fungal meningitis, endocarditis, and
resonance imaging scan of the brain (A) renal and chronic disseminated candidiasis (figures 4, 5, and
6). We know from several studies that patients who have

234 THE LANCET Infectious Diseases Vol 3 April 2003

For personal use. Only reproduce with permission from The Lancet Publishing Group.
Invasive fungal infections

Figure 6. A highly characteristic “air crescent” in a nodule in the left lung

at the level of the carina that is typical of invasive pulmonary aspergillosis,
Figure 5. CT scan of the thorax showing discrete nodules, which were a usually seen after the recovery of neutropenia, but occasionally in other
manifestation of chronic disseminated candidiasis after treatment for contexts.
acute leukaemia.

undergo CT scanning of the chest, ideally within 48 h of

leukaemia with invasive pulmonary aspergillosis are receipt of results (B). The positive predictive value of a
frequently diagnosed late because chest radiographs are positive respiratory-tract culture in solid-organ transplant
falsely negative (A). In these patients, conventional or spiral recipients is substantially lower than that in neutropenic
CT scanning of the chest should be done initially, followed allogeneic stem-cell-transplant patients. Large multicentre
by high-resolution CT of any abnormalities identified. Thin series have not been done, partly because of definition
slices (1 mm) should be taken, imaged with wide window difficulties, but figures quoted in the published research
settings (1500–2000 Hz), although thicker slices may be used vary from around 10% to more than 80%.30,45,71–73 In a US
for screening at a setting in which air and soft tissue can be series, the risk of invasive disease contemporaneously with
seen to best advantage (around 500 Hz).60 a positive culture of Aspergillus spp was 64% for allogeneic
In patients with abnormal chest radiographs, high- bone-marrow-transplant recipients and 17% in those who
resolution CT of the chest gives the best results. The had undergone solid-organ transplant.30
appearance of a so-called halo sign (an area of ground- The purpose of scanning is to identify rapidly patients
glass opacity) around a nodule or focal consolidation, who have invasive disease from those who are merely
especially seen during neutropenia, is highly characteristic colonised. Renal and heart transplant recipients have
of an invasive mould infection of the lung (A; figure 7). the best prognosis of all transplant recipients for invasive
The halo sign is present in more than 60% of patients, but aspergillosis, if the diagnosis is made. New pulmonary
for only 5–7 days during severe neutropenia, and opacities in these contexts are particularly useful.
occasionally in other patients such as allogeneic The differential diagnosis includes infection with Nocardia
haemopoietic stem-cell-transplant recipients.5,61–67 In
addition, chest radiography underestimates the extent of
disease (A).68
In patients with negative CT scans who have persistent
or recurrent fever, reimaging should be done with CT after
a short time (probably 7 days; B). Some patients with
invasive pulmonary aspergillosis require surgery
immediately to prevent catastrophic haemoptysis (B),5,69
and all those with pulmonary zygomycosis (A) if
technically possible.43 Failure rates of present treatment for
mould infections of the lung are high70 and alternative
treatments might be required. Radiological criteria are the
most important in determining therapeutic failure or lack
of response,47 although apparent progression during
neutropenia despite treatment can be consistent with a
delayed response.65
Figure 7. CT scan of the thorax showing a nodule and irregular
Transplant recipients consolidation adjacent to the pleura on the right with substantial
surrounding ground-glass consolidation that is typical of an obvious
All transplant recipients with a new positive culture of “halo” sign, and is mostly associated with invasive pulmonary
Aspergillus spp or other mould from any site should aspergillosis in neutropenic patients.

THE LANCET Infectious Diseases Vol 3 April 2003 235

For personal use. Only reproduce with permission from The Lancet Publishing Group.
Review Invasive fungal infections

spp, C neoformans, or Rhodococcus spp, community- Clinical management

acquired bacteria, viruses, and lymphoma. Subsequent Requests for investigations
investigations include bronchoscopy, serology (eg, for For sample processing requests in microbiology, histology,
cryptococcal antigen) and needle biopsy of the lung. and radiology to be prioritised correctly, request cards must
A sinus or nasal culture-yielding Aspergillus spp is unlikely include enough clinical information to show that the patient
to represent invasive disease in solid-organ-transplant is immunocompromised (A). Some hospitals may choose to
recipients, and this unusual situation should be exempted include an immunocompromised “check box” on forms.
from this recommendation. Patients with lesser degrees of suspected immune
dysfunction (such as those taking low-dose corticosteroids,
Immunocompromised patients with neurological antibodies to tumour necrosis factor, previous methotrexate
abnormalities treatment, but who are not neutropenic, etc) should be
All immunocompromised patients with new neurological classified as immunocompromised (B). Laboratories may
features (eg, change in mental status, seizure, stroke, need to set up new ways of scanning and interpreting clinical
persistent headache) or possible or proven meningitis data to take into account such changes.
should have a CT or magnetic resonance imaging scan
of the brain (A). Dissemination of an Aspergillus spp to the Candidaemia
brain is an occasional disastrous complication.70,74–76 The All patients with candidaemia should have central venous
triazoles itraconazole and voriconazole are probably catheters removed or replaced within 48 h of infection being
slightly better than amphotericin B for treatment of documented, and earlier if possible (A, evidence level 2,
cerebral aspergillosis.47,77,78 Unfortunately, the clinical panel 2), to lower mortality and shorten the duration of
presentation of cerebral aspergillosis is not distinctive infection.79–81 Potent antifungal drugs might lessen or remove
(with the exception of a seizure or new stroke) and can be this need in the future. Although only about 10% of cases of
insidious. Precise definition of precisely all the clinical candidaemia are thought to arise exogenously, seeding to the
features that should always prompt a scan in these complex catheter is common, if not universal, which provides an
immunocompromised patients is difficult. However, a infection focus that can remain subclinical. Peripheral
change in mental status not associated with drugs or other venous catheters do not seem to be a focus of infection (C3)
infectious process, although frequently subtle, is the and arterial lines should be changed as required (C3).
earliest clinical feature (A).74,75 Persistent new (but not All patients with candidaemia should be treated with a
necessarily continuous) headache is also a common early systemic antifungal agent at an appropriate dose, and
feature. A new seizure, abnormal neurological findings, or breakthrough fungaemia treated with an alternative agent
stroke are certainly indications for scanning (C). (unless all treatment is withdrawn—eg, in palliative care;
Dependent on the nature of the immunocompromising A2). Many cases of candidaemia were thought to be
conditions, the differential diagnosis of an abnormal brain transient and self-limiting, but studies now show there is no
scan may be broad.74 However in all transplant patients the means of distinguishing patients who have life-threatening
most frequent cerebral infectious complication is cerebral disease or who will develop focal complications such as
aspergillosis (figure 8). If coagulation parameters permit,
and the lesion identified is not too deep in the brain, Panel 6. Clinical management auditing standards in
biopsy or aspiration should be undertaken (A). systemic fungal disease (strength of recommendation)
Microscopy is as least as helpful in this context as analysis 1 All request cards for microbiology, histology, and radiology from
of bronchoalveolar lavage fluid, and the two approaches immunocompromised patients should be clearly identified as such in
should follow similar protocols (A). terms of patient information (A)
2 All patients with candidaemia should have central venous catheters
removed or replaced within 48 h of candidaemia being documented,
and preferably earlier, if technically possible, with the sole exception
being long-term lines that need a careful individual assessment (A2)
3 All patients with candidaemia should be treated with a systemic
antifungal agent at an appropriate dose, and breakthrough fungaemia
treated with an alternative agent (unless all treatment is withdrawn
[palliative care]) (A2)
4 All transplant recipients or profoundly neutropenic patients with:
l a new positive culture of aspergillus or other mould
l new pulmonary or cerebral abnormalities consistent with a fungal
l should be treated with a systemic antifungal agent at an appropriate
dose active against moulds within 6 h of the positive culture or
documentation of the pulmonary or cerebral abnormalities (A2)
5 All patients with cryptococcal meningitis should be treated initially with
amphotericin B deoxycholate >0·7 mg/kg/d or >4 mg/kg/d lipid-based
Figure 8. CT scan of the head showing a large contrast-enhancing amphotericin B with flucytosine 75–100 mg/kg/d (adjusted for renal
abscess in the left frontoparietal region, and substantial surrounding function) (A1)
oedema, typical of cerebral aspergillosis.

236 THE LANCET Infectious Diseases Vol 3 April 2003

For personal use. Only reproduce with permission from The Lancet Publishing Group.
Invasive fungal infections

ocular disease, endocarditis, osteomyelitis, and other foci, Early diagnosis and appropriate treatment of invasive
from those with transient candidaemia. Thus all patients mould infections are important in reducing mortality (A2).
with candidaemia should be treated with an appropriate Although some cultures represent colonisation or
antifungal drug (A2). Delays of more than 48 h in the contamination, many represent disease,30,45,71,100 with the one
starting of treatment worsens outcome.82 The intensity and common exception of lung transplant recipients.73 If the
duration of treatment may vary dependent on the disorder radiological differential diagnosis includes fungal infection,
and immune state of the patient, and the sites and species of treatment can be started (A2) and withdrawn if the culture
Candida involved. shows non-relevant colonisation or another established
Appropriate licensed systemic treatment includes only diagnosis (B3). The same is true of new radiological
fluconazole or amphotericin B,83–85 with or without abnormalities, although the specificity of some
flucytosine. The dose of fluconazole should not be less than abnormalities is much higher than for others.
400 mg daily (A2) and should be adjusted downwards in The only licensed choices for Aspergillus spp and other
patients with renal impairment (B3) and doubled in those on infections are amphotericin B, itraconazole, voriconazole,
continuous haemofiltration (B2).86 Candidaemia caused by and caspofungin. Only conventional amphotericin B and
C krusei and C glabrata should not be treated with voriconazole are licensed for primary treatment of invasive
fluconazole (A2). A few isolates of other species of Candida aspergillosis. Amphotericin B has been the broadest
(especially C albicans and C tropicalis) are fully or partly spectrum agent available and is active against A fumigatus,
resistant to fluconazole and are best treated with an A flavus, A niger, all the mucorales, Fusarium spp, and some
alternative drug (B2). Some data suggest that higher doses of other moulds, but it is not active against A terreus,
fluconazole (800 mg daily) may be more effective than Scedosporium spp, Paecilomyces spp, and a few other moulds.
standard doses of 400 mg daily, but this approach is not Itraconazole is active against all Aspergillus spp,
recommended unless the patient is extremely ill (B2).87,88 S apiospermum, and many other moulds, but is not active
There are concerns about use of fluconazole in profoundly against the mucorales, Fusarium spp, Scedosporium
neutropenic patients (C2), but these have been mostly prolificans, and some other rare moulds. Voriconazole has a
dispelled by careful matched-pair studies.89 slightly broader spectrum of activity against the moulds than
An alternative treatment for candidaemia is does itraconazole. By contrast, caspofungin and other
amphotericin B, but the dose required for a successful echinocandins have a narrow mould spectrum—probably
outcome is not known. Most authorities recommend at least Aspergillus spp only. Most infections are caused by
0·6 mg/kg daily (A1) and higher doses for very ill patients A fumigatus and A flavus, which are generally susceptible to
(B3) or those who have infections caused by all four drugs, although treatment outcomes are not good
C krusei, since this organism has poor susceptibility to overall.
amphotericin as well (B2).90,91 Lipid encapsulation probably The appropriate dose of a systemic antifungal drug with
increases the dose required for treatment based on limited activity against moulds, requires some interpretation. The
animal-model data, but there are conflicting clinical data on dose of amphotericin B should be 1·0 mg/kg daily (A2)8 or a
whether doses higher than 4·0 mg/kg do or do not yield a lipid-based preparation should be used with at least three
better outcome than doses lower than 3·0 mg/kg (C3).92,93 times this dose (3·0–10·0 mg/kg daily; C2). Itraconazole can
The indications for flucytosine in candidaemia are be given intravenously or orally, and loading doses should be
controversial because the drug has not been well studied91 given (A2). The standard dose intravenously is 200 mg once
and intravenous flucytosine is not available in the USA and daily, and orally is 200 mg twice daily for aspergillosis. The
some other countries. Our preferences are to use it only in standard treatment dose of itraconazole is also the
combination with fluconazole or amphotericin B in patients maintenance dose. Higher failure rates are associated with
with life-threatening infection (A2)94 and in those in whom lower doses of itraconazole (B3). Very large doses of
tissue penetration of the other agent may be poor—eg, itraconazole may be appropriate for cerebral aspergillosis.77
eye, urine, and meninges (B2).95–98 The UK registered dose Voriconazole, is better than amphotericin B (A1) and may
(200 mg/kg daily) is probably too high (A3) and is under provide the best alternative for invasive aspergillosis,47 but is
review. not active against mucorales (E2). Caspofungin may be a
Echinocandins (caspofungin having recently been useful alternative for invasive aspergillosis (A2).101
licensed) will probably have an important place in the
treatment of Candida spp infections.99 Cryptococcal meningitis
All patients with cryptococcal meningitis should be treated
Transplant recipients and neutropenic patients initially with amphotericin B deoxycholate at more than
All transplant recipients or severely neutropenic patients 0·7 mg/kg daily, or with lipid-based amphotericin B at more
who have a new positive culture of an Aspergillus spp or than 4·0 mg/kg daily with flucytosine 75–100 mg/kg
other mould, or new pulmonary or cerebral abnormalities daily (adjusted for renal function; A1).
consistent with a fungal infection should be treated with a Of all invasive fungal infections, the evidence base for the
systemic antifungal agent at an appropriate dose active treatment of cryptococcal meningitis is the best. The first
against moulds within 6 h of positive culture or randomised study in this disease showed that the addition of
documentation of the pulmonary or cerebral abnormalities flucytosine to a slightly lower dose of amphotericin B
(A2). improved response rates and lowered mortality (A1).102

THE LANCET Infectious Diseases Vol 3 April 2003 237

For personal use. Only reproduce with permission from The Lancet Publishing Group.
Review Invasive fungal infections

Search strategy and selection criteria amphotericin B formulation should be used. Poor results
were reported for Abelcet less than 3·0 mg/kg (D2),109 and
A consensus on the scope of the proposed standards was
achieved by the authors, based on their knowledge and
good responses with AmBisome 4·0 mg/kg daily. (B2).110 Few
experience. Additional advice was sought from colleagues who data are available for Amphocil. The penetration of
are specialists in mycology. For each standard, the relevant AmBisome into the brain is better than that with the other
published evidence was collected from filed reports and lipid-based amphotericin preparations and is equivalent to
searches in Medline. More than 70 individual searches were standard amphotericin B.111
done, many combining terms such as “candida” and
“catheter”, “microscopy” and “fungus”, “Calcafluor” and Conclusions
“mucorales”, etc. In subject areas for which large numbers of The absolute standards of care and proposed audit standards
reports exist, we have referred to those with the best evidence for patients with invasive fungal infection we describe
in favour of or against a recommendation. Reports published in
represent the current state of the art. Developments in the
languages other than English have not been included, and
books and book chapters were not accessed.
next few months or years may supersede these standards, such
as the routine advent of better serological and molecular
Other smaller studies comparing amphotericin B and diagnostic tests, better imaging methods, new antifungal
flucytosine with itraconazole or fluconazole showed better drugs, combination antifungal treatments, further emergence
response rates with the combination than with the azoles of antifungal resistance, and so on. Nonetheless, the pace of
alone (A1).103–106 A large randomised study of cryptococcal these future developments is unlikely to change many of these
meningitis in AIDS patients comparing amphotericin B with standards for the next 5 years.
and without flucytosine showed a better response rate with We could have proposed other non-absolute standards,
combination treatment than with amphotericin alone, but these can be difficult to audit and would have made
although not quite significantly so (A2),106 and a follow-up the review unwieldy. Clinicians, microbiologists,
study in responders done over 1 year, showed that relapse histopathologists, and radiologists experienced in this
was particularly associated with a lack of flucytosine in the specialty will, however, recognise gaps in the guidelines
first 2 weeks of treatment (A1).107 Thus, strong data support where the evidence base is less secure. These areas have not
the use of at least 2 weeks’ treatment with flucytosine been ignored by the British Society for Medical Mycology
combined with amphotericin B for the treatment of and, if the evidence base improves, will be the subject of
cryptococcal meningitis. further recommendations.
No good comparisons of the dose of amphotericin B
have been done. However, consistently reduced mortality
The recommendations contained in this paper have been endorsed
has been noted with higher doses of amphotericin B. A dose by the British Society for Medical Mycology committee (President,
of 0·7mg/kg daily is, therefore, recommended as a minimum Frank Odds). We thank Gillian Shankland, Paul McWhinney,
(B1), in combination with flucytosine. Higher doses of Barbara Isalska, Paul Bishop, Emyr Benbow, Paul Hulse, Helen
Williams at the Royal College of Pathologists, London, UK, and D Ash,
amphotericin B without flucytosine may be a reasonable at the Royal College of Radiologists, London, who all commented in
alternative in patients intolerant to flucytosine (B2).108 If detail on the standards.
toxic effects arise, or administration of amphotericin B is Conflicts of interest
impossible through a central venous catheter, a lipid-based None declared.

References 8 Stevens DA, Kan VL, Judson MA, et al. Practice correlation of in vitro amphotericin B susceptibility
1 Fraser DW, Ward JI, Ajello L, Plikaytis BD. guidelines for diseases caused by Aspergillus. Clin testing with outcome of in a murine model of
Aspergillosis and other systemic mycoses: the Infect Dis 2000; 30: 6696–709. Aspergillus infection. J Antimicrob Chemother 1999;
growing problem. JAMA 1979; 242: 1631–35. 9 Wingard JR, Merz WG, Rinaldi MG, Johnson TR, 45: 85–93.
2 Groll AH, Shah PM, Mentzel C, et al. Trends in the Karp JE, Saral R. Increase in Candida krusei 16 White TC, Marr KA, Bowden RA. Clinical, cellular,
postmortem epidemiology of invasive fungal infection among patients with bone marrow and molecular factors that contribute to antifungal
infections at a university hospital. J Infect 1996; 33: transplantation and neutropenia treated drug resistance. Clin Microbiol Rev 1998; 11:
23–32. prophylactically with fluconazole. N Engl J Med 382–402.
3 Rees JR, Pinner RW, Hajjeh RA, Brandt ME, 1991; 325: 1274–77. 17 Mosquera J, Warn PA, Rodriguez-Tudela JL, Denning
Reingold AL. The epidemiological features of 10 Vanden Bossche H, Dromer F, Improvisi I, DW. Treatment of Absidia corymbifera infection in
invasive mycotic infections in the San Francisco Bay Lozano-Chiu M, Rex JH, Sanglard D. Antifungal mice with amphotericin B and itraconazole. J
area, 1992–1993: results of population-based drug resistance in pathogenic fungi. Med Mycol Antimicrob Chemother 2001; 48: 583–86.
laboratory active surveillance. Clin Infect Dis 1998; 1998; 36 (suppl 1): 119–28. 18 Merz WG, Khazan U, Jabra-Rizk MA, Wu LC,
27: 1138–47. 11 Anaissie EJ, Karyotakis NC, Hachem R, Osterhout GJ, Lehmann PF. Strain delineation and
4 Yamazaki T, Kume H, Murase S, Yamashita E, Dignani MC, Rex JH, Paetznick V. Correlation epidemiology of Candida (Clavispora) lusitaniae.
Arisawa M. Epidemiology of visceral mycoses: between in vitro and in vivo activity of antifungal J Clin Microbiol 1992; 30: 449–54.
analysis of data in annual of the pathological autopsy agents against Candida species. J Infect Dis 1994; 170: 19 Finkelstein R, Reinhertz G, Hashman N, Merzbach
cases in Japan. J Clin Microbiol 1999; 37: 1732–38. 384–89. D. Outbreak of Candida tropicalis fungemia in a
5 Caillot D, Casasnovas O, Bernard A, et al. Improved 12 Denning DW, Baily GG, Hood SV. Azole resistance neonatal intensive care unit. Infect Control Hosp
management of invasive pulmonary aspergillosis in in Candida. Eur J Clin Microbiol Infect Dis 1997; 16: Epidemiol 1993; 14: 587–90.
neutropenic patients using early thoracic computed 261–80. 20 Welbel SF, McNeil MM, Kuykendall RJ, et al.
tomographic scan and surgery. J Clin Oncol 1997; 13 Lass-Flörl C, Kofler G, Kropshofer G, et al. In-vitro Candida parapsilosis bloodstream infections in
15: 139–47. testing of susceptibility to amphotericin B is a neonatal intensive care unit patients: epidemiologic
6 Barnes RA, Denning DW, Evans EGV, et al, for the reliable predictor of clinical outcome in invasive and laboratory confirmation of a common source
British Society for Antimicrobial Chemotherapy, aspergillosis. J Antimicrob Chemother 1998; 42: outbreak. Pediatr Infect Dis J 1996; 15: 998–1002.
and Denning DW, Evans EGV, Kibbler CC, et al, for 497–502. 21 Diekema DJ, Messer SA, Hollis RJ, Wenzel RP,
the British Society for Medical Mycology. Fungal 14 Sutton DA, Sanche SE, Revankar SG, Fothergill AW, Pfaller MA. An outbreak of Candida parapsilosis
infections: a survey of laboratory services for Rinaldi MG. In vitro amphotericin B resistance in prosthetic valve endocarditis. Diagn Microbiol Infect
diagnosis and treatment. CDR Rev 1996; 6: R69. clinical isolates of Aspergillus terreus, with a head-to- Dis 1997; 29: 147–53.
7 Kibbler C, Hamilton A. Medical mycology in the head comparison to voriconazole. J Clin Microbiol 22 Huang YC, Lin TY, Leu HS, Peng HL, Wu JH,
UK: time to stop the dry rot. Commun Dis Public 1999; 37: 2343–45. Chang HY. Outbreak of Candida parapsilosis
Health 2000; 3: 80–81. 15 Johnson E, Oakley KL, Radford S, et al. Lack of ungemia in neonatal intensive care units: clinical

238 THE LANCET Infectious Diseases Vol 3 April 2003

For personal use. Only reproduce with permission from The Lancet Publishing Group.
Invasive fungal infections

implications and genotyping analysis. Infection 1999; 47 Herbrecht R, Denning DW, Patterson TF, et al. rate: systematic review of the literature. Clin Infect
27: 97–102. Randomised comparison of voriconazole and Dis 2001; 32: 358–66.
23 Noskin GA, Lee J, Hacek DM, et al. Molecular typing amphotericin B in primary therapy of invasive 71 McWhinney PH, Kibbler CC, Hamon MD, et al.
for investigating an outbreak of Candida krusei. aspergillosis. N Engl J Med 2002; 347: 408–15. Progress in the diagnosis and management of
Diagn Microbiol Infect Dis 1996; 26: 117–23. 48 Segal BH, DeCarlo ES, Kwon-Chung KJ, Malech HL, aspergillosis in bone marrow transplantation: 13
24 Shin JH, Kook H, Shin DH, et al. Nosocomial cluster Gallin JI, Holland SM. Aspergillus nidulans infection years’ experience. Clin Infect Dis 1993; 17: 397–404.
of Candida lipolytica fungemia in pediatric patients. in chronic granulomatous disease. Medicine 1998; 72 Collins LA, Samore MH, Roberts MS, et al. Risk
Eur J Clin Microbiol Infect Dis 2000; 19: 344–49. 77: 345–54. factors for invasive fungal infections complicating
25 Aragao PA, Oshiro IC, Manrique EI, et al. Pichia 49 Moore CB, Sayers N, Mosquero J, Slaven J, Denning orthotopic liver transplantation. J Infect Dis 1994;
anomala outbreak in a nursery: exogenous source? DW. Antifungal drug resistance in Aspergillus. J 170: 644–52.
Pediatr Infect Dis J 2001; 20: 843–48. Infect 2000; 41: 203–20. 73 Cahill BC, Hibbs JR, Savik K, et al. Aspergillus airway
26 Marina NM, Flynn PM, Rivera GK, Hughes WT. 50 Hamilton JR, Noble A, Denning DW, Stevens DA. colonization and invasive disease after lung
Candida tropicalis and Candida albicans fungemia in Performance of cryptococcal antigen latex transplantation. Chest 1997; 112: 1160–64.
children with leukemia. Cancer 1991; 68: 594–99. agglutination kits on serum and cerebrospinal fluid 74 Hagensee ME, Bauwens JE, Kjos B, Bowden RA.
27 Voss A, Hollis RJ, Pfaller MA, Wenzel RP, specimens of AIDS patients before and after pronase Brain abscess following marrow transplantation:
Doebbeling BN. Investigation of the sequence of treatment. J Clin Microbiol 1991; 29: 333–39. experience at the Fred Hutchinson Cancer Research
colonization and candidemia in nonneutropenic 51 Mangham D, Gerding DN, Peterson LR, Sarosi GA. Center. 1984-1992. Clin Infect Dis 1994; 19: 402–08.
patients. J Clin Microbiol 1994; 32: 975–80. Fungal meningitis manifesting as hydrocephalus. 75 Pagano L, Ricci P, Montillo M, et al. Localization of
28 Rabkin JM, Oroloff SL, Corless CL, et al. Association Arch Intern Med 1983; 143: 728–31. aspergillosis to the central nervous system among
of fungal infection and increased mortality in liver 52 Mukerjee S, Casadevall A. Sensitivity of sandwich patients with acute leukemia: report of 14 cases. Clin
transplant recipients. Am J Surg 2000; 179: 426–30. enzyme-linked immunosorbent assay for Infect Dis 1996; 23: 628–30.
29 Nalesnik MA, Myerowitz RL, Jenkins R, Lenkey J, Cryptococcus neoformans polysaccharide antigen is 76 Denning DW. Therapeutic outcome of invasive
Herbert D. Significance of Aspergillus species isolated dependent on the isotypes of the capture and aspergillosis. Clin Infect Dis 1996; 23: 608–15.
from respiratory secretions in the diagnosis of detection antibodies. J Clin Microbiol 1995; 33: 77 Sanchez C, Mauri E, Dalmau D, Quintana S,
invasive pulmonary aspergillosis. J Clin Microbiol 765–68. Aparicio A, Garau J. Treatment of cerebral
1980; 11: 370–76. 53 Stevens DA. Coccidioidomycosis. N Engl J Med 1995; aspergillosis with itraconazole: do high doses
30 Perfect JR, Cox G M, Lee JY, et al, and the Mycoses 332: 1077–82. improve the prognosis? Clin Infect Dis 1995; 21:
Study Group. The impact of culture isolation of 54 Denning DW. Aspergillus meningitis. 1485–87.
Aspergillus species: a hospital-based survey of, 78 Denning DW, Ribaud P, Milpied N, et al. The
Aspergillosis. Clin Infect Dis 2001, 33: 1824–33. invasive (accessed Jan 20, 2002). efficacy and safety of voriconazole in the treatment
31 Kauffman CA, Vazquez JA, Sobel JD, et al. 55 Wheat J, Sarosi G, McKinsey D, et al. IDSA practice of acute invasive aspergillosis. Clin Infect Dis 2002;
Prospective multicenter surveillance study of guidelines for the management of patients with 34: 563–71.
funguria in hospitalized patients. Clin Infect Dis histoplasmosis. Clin Infect Dis 2000; 30: 688–95. 79 Eppes SC, Troutman JL, Gutman LT. Outcome of
2000; 30: 14–18. 56 Gottfredsson M, Perfect JR. Fungal meningitis. treatment of candidemia in children whose central
32 Berenguer J, Buck M, Witebsky F, Stock F, Pizzo PA, Semin Neurol 2000; 20: 307–22. catheters were removed or retained. Pediatr Infect
Walsh TJ. Lysis-centrifugation blood cultures in the 57 Ramos AM, Sales Ad, de Andrade MC, Bittencourt Dis J 1989; 8: 99–104.
detection of tissue-proven invasive candidiasis: JF, Ramos CC. A simple method for detecting 80 Dato VM, Dajani AS. Candidemia in children with
disseminated versus single-organ infection. Diagn subcutaneous phaeohyphomycosis with light- central venous catheters: role of catheter removal
Microbiol Infect Dis 1993; 17: 103–09. colored fungi: a study of eight cases. Am J Surg and amphotericin B therapy. Pediatr Infect Dis J
33 Washington JA. Collection, transport, and Pathol 1995; 19: 109–14. 1990; 9: 309–14.
processing of blood cultures. Clin Lab Med 1994; 14: 58 Husain AN, Siddiqui MT, Montoya A, 81 Rex JH, Bennett JE, Sugar AM, et al. Intravascular
59–68. Chandrasekhar AJ, Garrity ER. Post-lung transplant catheter exchange and duration of candidemia.
34 Fuller DD, Davis TE Jr, Denys GA, York MK. biopsies: an 8-year Loyola experience. Mod Pathol NIAID Mycoses Study Group and the Candidemia
Evaluation of BACTEC MYCO/F lytic medium for 1996; 9: 126–32. Study Group. Clin Infect Dis 1995; 21: 994–96.
recovery of mycobacteria, fungi, and bacteria from 59 Fons A, Garcia-de-Lomas J, Nogueira JM, Buesa FJ, 82 Nolla-Salas J, Sitges-Serra A, Leon-Gil C, et al.
blood. J Clin Microbiol 2001; 39: 2933–36. Gimeno C. Histopathology of experimental Candidaemia in non-neutropenic critically ill
35 Kahn FW, Jones JM, England DM. The role of Aspergillus fumigatus keratitis. Mycopathologia 1988; patients: analysis of prognostic factors and
bronchoalveolar lavage in the diagnosis of invasive 101: 129–31. assessment of systemic antifungal therapy: study
pulmonary aspergillosis. Am J Clin Pathol 1986; 86: 60 Berger LA. Imaging in the diagnosis of infections in group of fungal infection in the ICU. Intensive Care
518–23. immunocompromised patients. Curr Opinion Infect Med 1997; 23: 23–30.
36 Levy H, Horak DA, Tegtmeier BR, Yokota SB, Dis 1998; 11: 431–36. 83 Rex JH, Bennett JE, Sugar AM, et al. A randomized
Forman SJ. The value of bronchoalveolar lavage and 61 Kuhlman, JE, Fishman EK, Siegelman SS. Invasive trial comparing fluconazole with amphotericin B for
bronchial washings in the diagnosis of invasive pulmonary aspergillosis in acute leukemia: the treatment of candidemia in patients without
pulmonary aspergillosis. Respir Med 1992; 86: characteristic findings on CT, the CT halo sign, and neutropenia: Candidemia Study Group and the
243–48. the role of CT in early diagnosis. Radiology 1985; NIAID. N Engl J Med 1994; 331: 1325–30.
37 Fischler DF, Hall GS, Gordon S, Stoler MH, Nunez 157: 611–14. 84 Anaissie EJ, Darouiche RO, Abi-Said D, et al.
C. Aspergillus in cytology specimens: a review of 45 62 Moro M, Galvin JR, Barloon TJ, Gingrich RD, Management of invasive candidal infections: results
specimens from 36 patients. Diagn Cytopathol 1997; Stanford W. Fungal pulmonary infections after bone of a prospective, randomized, multicenter study of
16: 26–30. marrow transplantation: evaluation with fluconazole versus amphotericin B and review of the
38 Yuen KY, Woo PC, Ip MS, et al. Stage-specific radiography and CT. Radiology 1991; 178: 721–26. literature. Clin Infect Dis 1996; 23: 964–72.
manifestations of mold infections in bone marrow 63 Primack SL, Hartman TE, Lee, Müller KS. 85 Phillips P, Shafran S, Garber G, et al. Multicenter
transplant recipients: risk factors and clinical Pulmonary nodules and the CT halo sign. Radiology randomized trial of fluconazole versus amphotericin
significance of positive concentrated smears. Clin 1994; 190: 513–15. B for treatment of candidemia in non-neutropenic
Infect Dis 1997; 25: 37–42. 64 Blum U, Windfuhr M, Buitrago-Tellez C, Sigmund patients: Canadian Candidemia Study Group.
39 Pagano L, Pagliari G, Basso A, et al. The role of G, Herbst EW, Langer M. Invasive pulmonary Eur J Clin Microbiol Infect Dis 1997; 16: 337–45.
bronchoalveolar lavage in the microbiological aspergillosis: MRI, CT, and plain radiographic 86 Muhl E, Martens T, Iven H, Rob P, Bruch HP.
diagnosis of pneumonia in patients with findings and their contribution for early diagnosis. Influence of continuous veno-venous
haematological malignancies. Ann Intern Med 1997; Chest 1994; 106: 1156–67. haemodiafiltration and continuous veno-venous
29: 535–40. 65 Caillot D, Couaillier J-F, Bernard A, et al. Increasing haemofiltration on the pharmacokinetics of
40 Chander J, Chakrabarti A, Sharma A, Saini JS, volume and changing characteristics of IPA on fluconazole. Eur J Clin Pharmacol 2000; 56: 671–78.
Panigarhi D. Evaluation of Calcofluor staining in the sequential thoracic CT scans in neutropenic patients. 87 Graninger W, Presterl E, Schneeweiss B, Teleky B,
diagnosis of fungal corneal ulcer. Mycoses 1993; 36: J Clin Oncol 2001; 19: 253–59. Georgopoulos A Treatment of Candida albicans
243–45. 66 Heussel CP, Kauczor HU, Heussel GE, et al. fungaemia with fluconazole. J Infect 1993; 26:
41 Andreas S, Heindl S, Wattky C, Moller K, Ruchel R. Pneumonia in febrile neutropenic patients and in 133–46.
Diagnosis of pulmonary aspergillosis using optical bone marrow and blood stem-cell transplant 88 Rex JH, Pappas PG, Karchmer AW, et al.
brighteners. Eur Respir J 2000; 15: 407–11. recipients: use of high resolution computed Randomised and blinded multicenter trial of high-
42 Monod M, Jaccoud S, Stirnimann R, et al. tomography. Clin Oncol 1999; 17: 796–805. dose fluconazole (F) + placebo (P) vs F +
Economical microscope configuration for direct 67 Yeghen T, Kibbler CC, Prentice HG, et al. amphotericin B (A) as treatment of candidemia in
mycological examination with fluorescence in Management of invasive pulmonary aspergillosis in non-neutropenic patients. ICAAC; 2001 Sept 22–25;
dermatology. Dermatology 2000; 201: 246–48. hematology patients: a review of 87 consecutive cases Chicago, USA: abstr J681a, p 378.
43 Tedder M, Spratt J A, Anstadt M P, Hedge SS, at a single institution. Clin Infect Dis 2000; 31: 89 Anaissie EJ, Vartivarian SE, Abi-Said D, et al.
Tedder SD, Lowe JE. Pulmonary mucormycosis: 859–68. Fluconazole versus amphotericin B in the treatment
results of medical and surgical therapy. Ann Thorac 68 Denning DW, Marinus A, Cohen J, et al, and the of hematogenous candidiasis: a matched cohort
Surg 1994; 57: 1044–50. EORTC Invasive Fungal Infections Cooperative study. Am J Med 1996; 101: 170–76.
44 Kovacs JA, Gill VJ, Meshnick S, Masur H. New Group. An EORTC multicentre prospective survey 90 Edwards JE Jr, Bodey GP, Bowden RA, et al.
insights into transmission, diagnosis and drug of invasive aspergillosis in cancer patients: diagnosis International conference for the development of a
treatment of Pneumocystis carinii pneumonia. JAMA and therapeutic outcome. J Infect 1998; 37: 173–80. consensus on the management and prevention of
2001; 286: 2450–60. 69 Pagano L, Ricci P, Nosari A, et al, for the Gimema severe candidal infections. Clin Infect Dis 1997; 25:
45 Horvath JA, Dummer, S. The use of respiratory-tract Infection Program (Gruppo Italiano Malattie 43–59.
cultures in the diagnosis of invasive pulmonary Ematologiche dell’Adulto). Fatal haemoptysis in 91 Rex JH, Walsh TJ, Sobel JD, et al, for the Infectious
aspergillosis. Am J Med 1996; 100: 171–78. pulmonary filamentous mycosis: an underevaluated Diseases Society of America. Practice guidelines for
46 Johnson E, Oakley KL, Radford S, et al. Lack of cause of death in patients with acute leukaemia in the treatment of candidiasis. Clin Infect Dis 2000; 30:
correlation of in vitro amphotericin B susceptibility haematological complete remission—a retrospective 662–78.
testing with outcome of in a murine model of study and review of the literature. Br J Haematol 92 Oppenheim BA, Herbrecht R, Kusne S. The safety
Aspergillus infection. J Antimicrob Chemother 2000; 1995; 89: 500–05. and efficacy of amphotericin B colloidal dispersion
45: 85–93. 70 Lin S, Schranz J, Teutsch S. Aspergillosis case-fatality in the treatment of invasive mycoses. Clin Infect Dis

THE LANCET Infectious Diseases Vol 3 April 2003 239

For personal use. Only reproduce with permission from The Lancet Publishing Group.
Review Invasive fungal infections

1995; 21: 1145–53. noncomparative study to evaluate safety and efficacy National Institute of Allergy and Infectious Diseases
93 Linden P, Lee L, Walsh TJ. Retrospective analysis of of caspofungin (CAS) in adults with invasive Mycoses Study Group and AIDS Clinical Trials
the dosage of amphotericin B lipid complex for the aspergillosis (IA) refractory (R) or intolerant (I) Group. N Engl J Med 1997; 337: 15–21.
treatment of invasive fungal infections. to amphotericin B (AMB), AMB lipid formulations 107 Saag MS, Cloud GA, Graybill JR, et al, for the
Pharmacotherapy 1999; 19: 1261–68. (lipid AMB), or azoles. Interscience Conference National Institute of Allergy and Infectious Diseases
94 Vincent JL, Anaissie E, Bruining H, et al. on Antimicrobial Agents and Chemotherapy; Mycoses Study Group. A comparison of itraconazole
Epidemiology, diagnosis and treatment of systemic Toronto, USA; 2000 Sept 16–20; abstr 1103, p 371. versus fluconazole as maintenance therapy for AIDS-
Candida infection in surgical patients under 102 Bennett JE, Dismukes WE, Duma RJ, et al. associated cryptococcal meningitis. Clin Infect Dis
intensive care. Intensive Care Med 1998; 24: 206–16. A comparison of amphotericin B alone and 1999; 28: 291–96.
95 Walsh A, Haft DA, Miller MH, Loran MR, Friedman combined with flucytosine in the treatment of 108 Hsieh SM, Hung CC, Chen MY, Hsueh PR, Chang
AH. Ocular penetration of 5-fluorocytosine. Invest cryptococcal meningitis. N Engl J Med 1979; 301: SC. Efficacy and adverse effects of higher dose
Ophthalmol Vis Sci 1978; 17: 691–94. 126–31. amphotericin B monotherapy for cryptococcal
96 Harder EJ, Hermans PE. Treatment of fungal 103 Larsen RA, Leal MA, Chan LS. Fluconazole meningitis in patients with advanced HIV infection.
infections with flucytosine. Arch Intern Med 1975; compared with amphotericin B plus flucytosine for J Microbiol Immunol Infect 1998; 31: 233–39.
135: 231–37. cryptococcal meningitis in AIDS: a randomized trial. 109 Sharkey PK, Graybill JR, Johnson ES, et al.
97 Wise GJ, Kozinn PJ, Goldberg P. Flucytosine in the Ann Intern Med 1990; 113: 183–87. Amphotericin B lipid complex compared with
management of genitourinary candidiasis: 5 years of 104 Saag MS, Powderly WG, Cloud GA, et al, for the amphotericin B in the treatment of cryptococcal
experience. J Urol 1980; 124: 70–72. NIAID Mycoses Study Group and the AIDS Clinical meningitis in patients with AIDS. Clin Infect Dis
98 Rowen JL, Tate JM, for the Neonatal Candidiasis Trials Group. Comparison of amphotericin B with 1996; 22: 315–21.
Study Group. Management of neonatal candidiasis. fluconazole in the treatment of acute AIDS- 110 Leenders AC, Reiss P, Portegies P, et al. Liposomal
Pediatr Infect Dis J 1998; 17: 1007–11. associated cryptococcal meningitis. N Engl J Med amphotericin B (AmBisome) compared with
99 Mora-Duarte J, Betts R, Rotstein C, et al. 1992; 326: 83–89. amphotericin B both followed by oral fluconazole in
Comparison of caspofungin and amphotericin B for 105 de Gans J, Portegies P, Tiessens G, et al. Itraconazole the treatment of AIDS-associated cryptococcal
invasive candidiasis. N Engl J Med 2002; 347: 2020. compared with amphotericin B plus flucytosine in meningitis. AIDS 1997; 11: 1463–71.
100 Greub G, Bille J. Aspergillus species isolated from AIDS patients with cryptococcal meningitis. 111 Groll AH, Giri N, Petraitis V, et al. Comparative
clinical specimens: suggested clinical and AIDS 1992; 6: 185–90. efficacy and distribution of lipid formulations of
microbiological criteria to determine significance. 106 van der Horst CM, Saag MS, Cloud GA, et al. amphotericin B in experimental Candida albicans
Clin Microbiol Infect 1998; 4: 710–16. Treatment of cryptococcal meningitis associated infection of the central nervous system. J Infect Dis
101 Maertens J, Raad I, Sable CA, et al. Multicenter, with the acquired immunodeficiency syndrome. 2000; 182: 274–82.

240 THE LANCET Infectious Diseases Vol 3 April 2003

For personal use. Only reproduce with permission from The Lancet Publishing Group.